Biochemical Society Transactions

RNA UK 2012

Comparison of the yeast and human nuclear exosome complexes

Katherine E. Sloan, Claudia Schneider, Nicholas J. Watkins


Most RNAs in eukaryotic cells are produced as precursors that undergo processing at the 3′ and/or 5′ end to generate the mature transcript. In addition, many transcripts are degraded not only as part of normal recycling, but also when recognized as aberrant by the RNA surveillance machinery. The exosome, a conserved multiprotein complex containing two nucleases, is involved in both the 3′ processing and the turnover of many RNAs in the cell. A series of factors, including the TRAMP (Trf4–Air2–Mtr4 polyadenylation) complex, Mpp6 and Rrp47, help to define the targets to be processed and/or degraded and assist in exosome function. The majority of the data on the exosome and RNA maturation/decay have been derived from work performed in the yeast Saccharomyces cerevisiae. In the present paper, we provide an overview of the exosome and its role in RNA processing/degradation and discuss important new insights into exosome composition and function in human cells.

  • exonuclease
  • exosome
  • RNA degradation
  • RNA processing
  • RNA surveillance

Exosome functions

The exosome is a multiprotein nuclease complex that functions in RNA processing and turnover. This complex was first discovered in Saccharomyces cerevisiae and has been extensively studied in this organism, whereas the functions of the human exosome are currently less well characterized [1]. The exosome is responsible for the 3′→5′ exonuclease processing and/or degradation of a wide range of RNA substrates both in the nucleus and cytoplasm (Figure 1) (reviewed in [25] and references therein). The 3′ processing of many non-coding RNAs in the nucleus, including the snoRNAs (small nucleolar RNAs), snRNAs (small nuclear RNAs) and 5.8S rRNA, is mediated by the exosome [2,3]. In the nucleus, the exosome also participates in the recycling of the by-products of RNA processing such as fragments of pre-rRNA released by serial endonucleolytic cleavages and intron lariats produced by pre-mRNA splicing (reviewed in [2,3,5]). Further, the exosome is central to RNA surveillance and quality control; defective sn/snoRNAs, pre-rRNAs, tRNAs or mRNAs, which are not correctly processed, spliced or assembled into ribonucleoprotein complexes are targeted for degradation by the exosome (reviewed in [4]). Other RNA polymerase II transcripts such as CUTs (cryptic unstable transcripts), which regulate gene expression in yeast, and PROMPTs (promoter upstream transcripts), which are found in human cells, are also turned over by the exosome in the nucleus (reviewed in [2]).

Figure 1 Schematic overview of exosome substrates in eukaryotes

The core exosome represented by blue/purple ovals and various exosome substrates in the nucleus and cytoplasm are shown. Cofactors that participate in the pathways of RNA processing/turnover are named along the corresponding arrows. ARE, AU-rich element; ARE-BP, ARE-binding protein; NGD, no-go decay; NMD, nonsense-mediated decay; NSD, non-stop decay; TUTase, terminal uridylyltransferase.

Eukaryotic core exosome structure

Crystal structures of yeast and human exosomes show that nine exosomal subunits are organized into a hexameric ring and cap [6] (Figure 2). The PH domain subunits Rrp41, Rrp46, Mtr3, Rrp42, Rrp43 (in humans also known as OIP2) and Rrp45 (PM/Scl75) assemble to form a doughnut ring-shape that is stabilized by a trimeric cap formed by three KH/S1 domain proteins, Rrp4, Rrp40 and Csl4, which assist in RNA binding and substrate recognition. The cap proteins are clustered at the entrance to the channel on the one side of the ring, enabling substrates to be threaded through the central pore [6,7]. The core structure is reminiscent of the bacterial PNPase (polynucleotide phosphorylase) and archaeal exosome. However, unlike the archaeal and bacterial complexes, the eukaryotic nine subunit exosome core does not possess nuclease activity [6,8,9]. Many of the proteins of the human exosome complex were first identified as autoantigens that are found in the autoimmune disorders PM (polymyositis), Scl (scleroderma) and the PM/Scl overlap syndrome [10,11]. Further, the core exosome subunit, RRP46, is significantly up-regulated in patients with chronic myeloid leukaemia and RRP46 may be a tumour-related antigen that stimulates autoimmune responses in cancer [12,13].

Figure 2 Schematic model of TRAMP-stimulated RNA degradation by the yeast exosome

The PH domain proteins form a hexameric ring, from which Mtr3 and Rrp42 have been omitted. Threading of a structured RNA (broken line) through the exosome core to the exonuclease site of Rrp44 by Mtr4 and the TRAMP complex is shown. Figure based on Figure 3 in [7].

Yeast catalytic subunits

The activity of the eukaryotic core exosome is derived by association with two nucleases, Rrp44/Dis3 and Rrp6. In yeast, Rrp44/Dis3 is considered the tenth core exosome subunit as it is constitutively associated with exosome complexes in both the nucleus and cytoplasm and is essential for cell viability [6,8,14,15]. Rrp44/Dis3 is a highly processive exonuclease that is homologous with Escherichia coli RNaseR, a member of the RNase II family of hydrolytic 3′→5′ exonucleases [16]. Rrp44/Dis3 also harbours endonuclease activity in its N-terminal PIN domain [1719]. This PIN domain is also required for the association of Rrp44/Dis3 with the lower face of the core exosome, on the opposite end to the trimeric cap [7,19] (Figure 2). Single-stranded substrates are thought to pass though the central pore of the core hexameric ring to the exonucleolytic catalytic centre of Rrp44/Dis3 (Figure 2). The endo- and exo-nuclease activities of Rrp44/Dis3 are proposed to be co-ordinated, enabling the degradation of large, structured RNAs [1719]. Interestingly, the association of Rrp44/Dis3 with the exosome core regulates its exonuclease activity [6,8].

Rrp6 (PM/Scl100 in humans) is a member of the DEDD superfamily of 3′→5′ ribonucleases and deoxyribonucleases and is homologous with E. coli RNase D [20,21]. In contrast with Rrp44/Dis3, yeast Rrp6 has distributive exonuclease activity and is only capable of degrading unstructured RNA substrates [22]. Rrp6 is only associated with nuclear exosome complexes and it is not yet clear how this interaction is mediated. Different models are proposed in which Rrp6 binds to either the KH domain cap proteins or to the PH domain hexameric core proteins [23,24]. In the context of nuclear exosome complexes, Rrp44/Dis3 and Rrp6 can work together to process RNA substrates, e.g. 5.8S rRNA. Interestingly, Rrp6 is also able to function independently of the core exosome in the 3′ processing of snoRNAs and 5.8S rRNA [25].

Human catalytic subunits

Several recent papers have highlighted significant differences in the exonuclease and cofactor composition of yeast and human exosome complexes in different subcellular compartments. Although only one form of Rrp44/Dis3 is present in yeast, three different human homologues have been identified: DIS3, DIS3L1 and DIS3L2. Of these, only DIS3 and DIS3L1 appear to be associated with the core exosome and it seems that these proteins are less tightly associated with the core exosome than yeast Rrp44/Dis3 is [26,27]. The PIN domain of DIS3L2 is poorly conserved and as this domain is required for Rrp44/Dis3 association with the yeast exosome, this is likely to explain why DIS3L2 is not found in purified exosome complexes [27]. The three DIS3 isoforms have distinct subcellular localizations, with DIS3 being predominantly nucleoplasmic but excluded from nucleoli, whereas DIS3L1 and DIS3L2 are found exclusively in the cytoplasm [27] (Figure 3). Although all three human DIS3 proteins have 3′→5′ exonuclease activity in vitro, only DIS3 was also found to have endonuclease activity like yeast Rrp44. This may be reflective of the higher proportion of structured RNA substrates in the nucleus [27]. Exosome-associated RNA processing activities of both DIS3 and DIS3L1 have been demonstrated in vivo [26,27]. RNAi (RNA interference)-mediated depletion of DIS3, but not DIS3L1 or DIS3L2 causes accumulation of 3′ extended precursors of 5.8S rRNA. DIS3 depletion also leads to the stabilization of PROMPTs in the nucleus [27,28]. Depletion of DIS3L1 causes significant stabilization of c-Myc and c-Fos mRNAs [27] in the cytoplasm, and the accumulation of oligo(A)-tailed rRNA degradation fragments [26], which are both normally degraded by the exosome. In addition to functions in RNA processing, both yeast Rrp44/Dis3 and human DIS3L2 are important for microtubule structure and mitotic progression [28,29]. Further, mutations in DIS3L2 are linked to a genetic disease, Perlman syndrome, and lead to an increased frequency of Wilms' tumours [28].

Figure 3 Model of subcellular localization of different forms of the human exosome and cofactors

The core exosome is shown in blue and purple. The PIN domains of DIS3 (active-small ‘pacman’) and DIS3L1 (inactive-small circle) are shown.

As in yeast, human RRP6 or PM/Scl100 possesses distributive exonucleolytic activity, but human RRP6 is able to degrade more structured substrates than its yeast counterpart [22]. This is probably because of a more open conformation of the catalytic site in the human protein and may suggest that human RRP6 is able to process different substrates to its yeast counterpart. Unlike yeast Rrp6, which is strictly nuclear, human RRP6 is highly concentrated in nucleoli and also found, at lower levels, in the nucleoplasm and cytoplasm [27,30] (Figure 3). This implies that the majority of the exosome activity in the nucleolus of human cells is provided by RRP6 and suggests that RRP6 may also have cytoplasmic substrates in higher eukaryotes.

Yeast nuclear cofactors

Given the large range of potential substrates encountered by the exosome, how the complex is targeted to particular substrates remains a major question. It is also not yet clear how the exosome is able to distinguish between substrates that are to be processed and those that are to be degraded. It is thought that this specificity is achieved by association of the exosome with a number of cofactors, which are important for substrate identification (Figure 1). In addition, some of these accessory proteins regulate the activity of the exosome in vivo.

A well-characterized cofactor of the yeast nuclear exosome is the TRAMP (Trf4/5–Air1/2–Mtr4 polyadenylation) complex [3133]. The TRAMP complex functions in RNA processing and surveillance both by recruiting the exosome to its RNA substrates through addition of short poly(A) tails and also by stimulating exonucleolytic degradation by the exosome. The TRAMP complex is composed of a non-canonical poly(A) polymerase (either Trf4 or Trf5), a putative RNA-binding protein (either Air1 or Air2) and the RNA helicase Mtr4 [3133]. Two different TRAMP complexes have been identified in yeast: Trf4–Air1/2–Mtr4 and Trf5–Air1–Mtr4 [3133]. Trf4/5 are homologous distributive polymerases that add short poly(A) tails to RNA substrates in the context of the TRAMP complex. The RNA helicase Mtr4 modulates the relative kinetics of adenylation and TRAMP dissociation, thereby restricting the number of adenosines added to approximately 4 or 5 [34,35]. Trf5 activity appears to be restricted to the nucleolus, whereas Trf4 apparently only functions in the nucleus [4,33,36]. The Air1/2 proteins, which contain five zinc knuckle motifs, are homologous and functionally redundant. The Air proteins are predicted to be important for substrate binding and specificity of the TRAMP complex.

The TRAMP complex stimulates the exonuclease activity of Rrp6 and Rrp44 in vitro [9,32,37]. The helicase activity of Mtr4 is proposed to unwind RNA secondary structures and displace RNA-associated proteins ahead of exonucleolytic processing by the exosome (Figure 2). A model is proposed in which the ATPase activity of Mtr4 drives RNA substrates through the helicase domain, analogous to the mechanism used by the proteasome, and as these RNAs exit, they are directly channelled towards Rrp6 and through the core exosome to Rrp44/Dis3 for processing [38,39]. The crystal structure of Mtr4 revealed a novel ‘arch’ or ‘stalk’ domain that is not found in canonical helicases [38,39]. This structure is essential for the function of Mtr4 in the 3′-end processing of 5.8S rRNA and disruption causes accumulation of 3′ extended RNA fragments similar to those seen upon deletion of Rrp6 [38]. Mtr4 also functions as a cofactor of the exosome independently of the TRAMP complex, and as such, is involved in processing of the 3′-end of 5.8S rRNA in yeast and plant cells [40,41].

Other proteins that co-purify with yeast nuclear exosome complexes, primarily through interactions with Rrp6, are Rrp47 (C1D in humans) and the M-phase phosphoprotein 6, Mpp6 [42,43]. Both of these nuclear proteins are involved in RNA binding and they have been shown to preferentially interact with structured and pyrimidine-rich sequences respectively [44]. Rrp47 contains an unusual RNA-binding domain called a Sas10/C1D domain, which is also found in pre-rRNA binding factors, Utp3 and Lcp5 [45]. Rrp47 interacts with the PMC2NT domain of Rrp6, suggesting that it may function as an adaptor recruiting the Rrp6-containing exosome to its substrates [43,44]. Depletion of Rrp47 causes RNA processing defects similar to those caused by lack of Rrp6, and both Rrp47 and Mpp6 are required for 3′ processing of 5.8S rRNA in yeast [42,43]. In yeast, Mpp6 is also involved in exosomal degradation of unstable non-coding RNAs produced by RNA polymerase II [42].

The Nrd1–Nab3 complex, which also contains the helicase Sen1, co-ordinates the termination of transcription of various RNA polymerase II transcripts, including pre-snRNAs, pre-snoRNAs and CUTs, with the initiation of either their maturation or degradation by the exosome and TRAMP complexes [46]. This complex is also involved in surveillance of RNA polymerase III transcripts including pre-tRNAs [35].

Cofactors of the human exosome

Recent data have identified human homologues of the proteins found in the yeast TRAMP complexes, suggesting that these complexes are conserved in higher eukaryotes. Human MTR4, has been characterized as a nuclear protein that is strongly concentrated in nucleoli and is involved in 3′ processing of human 5.8S rRNA [47]. Owing to the presence of conserved zinc knuckles and IWRXY motifs, ZCCHC7 and ZCCHC8 are proposed homologues of the yeast Air proteins [48,49]. Based on the sequence homology, the human counterparts of yeast Trf4 and Trf5 have been identified as PAPD5/hTRF4–2 and PAPD7 respectively. ZCCHC7 and PAPD5/hTRF4–2 localize to the nucleolus, interact directly and have been shown to be precipitated by both MTR4 and RRP6, implying that they are associated with the core exosome [48,49]. RNAi depletion of either ZCCHC7 or PAPD5/hTRF4–2 leads to a decrease in the accumulation of polyadenylated 5′-ETS pre-rRNA fragments that are normally turned over by the nuclear exosome, confirming that these proteins are functional cofactors of the human exosome [49]. Analogous to the nuclear TRAMP complex found in yeast, another protein complex has been identified in the nucleus of human cells called the NEXT (nuclear exosome targeting) complex [49]. In addition to MTR4, this complex also contains ZCCHC8 and RBM7 and is involved in the degradation of PROMPTs, but does not appear to participate in 3′ processing of 5.8S rRNA [49].

C1D (Rrp47) and MPP6 are also conserved in higher eukaryotes and, like the yeast proteins, involved in 3′ processing of 5.8S [47,50]. It is suggested that there may be additional cofactors of the human nuclear exosome and these include the zinc knuckle proteins, ZFC3H1, ZCH18 and ARS2 [49]. Surprisingly, Nrd1 and Nab3 have yet to be identified in humans, suggesting that the function of these factors may not be required in higher eukaryotes.


In contrast with the yeast cytoplasmic (Rrp44) and nuclear (Rrp44 and Rrp6) exosomes, the composition of active human exosomes in different subcellular compartments is significantly more complex (Figure 3). This increased complexity is also reflected in the distribution and interactions of the TRAMP-like cofactors. It is likely that this variability provides both greater versatility and specificity, which are required to deal with the much larger number, and more diverse range, of transcripts in the human genome. Characterization of these different human exosome complexes, their regulators/adaptors and their substrate interactions will be key to understand how transcripts from such a complex genome are regulated and processed.


N.J.W. was supported by a grant from the Wellcome Trust and C.S. was supported by a Royal Society University Research Fellowship.


  • RNA UK 2012: An Independent Meeting held at The Burnside Hotel, Bowness-on-Windermere, Cumbria, U.K., 20–22 January 2012. Organized and Edited by Raymond O'Keefe and Mark Ashe (Manchester, U.K.).

Abbreviations: CUT, cryptic unstable transcript; PM, polymyositis; PROMPT, promoter upstream transcript; RNAi, RNA interference; Scl, scleroderma; snRNA, small nuclear RNA; snoRNA, small nucleolar RNA; TRAMP, Trf4–Air2–Mtr4 polyadenylation


View Abstract